Sanjaya Lall Memorial Lecture 2024
Summary
TLDRÎn această prelegere, profesorul Darren Acemoglu abordează subiectul provocărilor și oportunităților generate de noile tehnologii, precum inteligența artificială (AI), asupra economiei și societății. El se concentrează pe două aspecte esențiale: rolul relațiilor de putere și natura tehnologică a automatizării în determinarea impactului economic al inovațiilor. Prin exemple istorice, precum Revoluția Industrială Britanică și evoluția fabricilor auto din secolul XX, Acemoglu ilustrează modul în care tehnologiile noi au generat atât progres economic, cât și provocări legate de distribuția inechitabilă a veniturilor. Importanța instituțiilor democratice și a inovării direcționate responsabil este subliniată ca fiind esențială pentru a asigura că beneficiile tehnologiilor sunt împărțite echitabil. Profesorul avertizează asupra consecințelor unei automatisări necontrolate și consideră că o participare mai activă a societății civile și reglementări adecvate sunt necesare pentru a naviga cu succes era digitală.
Takeaways
- 🤖 Tehnologia poate crește eficiența producției, dar nu garantează distribuția echitabilă a prosperității.
- ⚖️ Relațiile de putere influențează modul în care beneficiile tehnologice sunt distribuite în societate.
- 📉 Automatizarea excesivă poate reduce valoarea marginală a muncii și salariile reale.
- ⚙️ Istoria arată că inovațiile nu au condus întotdeauna la creșterea nivelului de trai pentru toți.
- 🔄 Nevoia de instituții democratice și sindicalizare pentru un echilibru economic echitabil.
- 💡 Tehnologiile pot fi utilizate pentru a îmbunătăți productivitatea umană dacă sunt direcționate corect.
- 🏭 Exemplele istorice arată cum tehnologia poate perpetua inegalitatea dacă nu este gestionată corect.
- 💬 AI și digitalizare: posibile atât pentru automatizare, cât și pentru îmbunătățirea abilităților umane.
- 🌍 Impactul global al AI și cum poate schimba diviziunea muncii între țări.
- 📊 Importanța reglementărilor și a societății civile în orientarea progresului tehnologic.
Timeline
- 00:00:00 - 00:05:00
Evenimentul este deschis cu un discurs de bun venit și introducerea lui Darren Asoglu, un renumit profesor de economie la MIT și vizitator la Oxford. Amfitrionul Andrew Steven oferă un context despre conferința memorială Sanaya La și contribuțiile notabile ale profesorului Zan Al.
- 00:05:00 - 00:10:00
Darren își prezintă lucrarea din cartea sa, concentrându-se pe impactul tehnologiilor asupra societății și economiei. Subiectul central este cine beneficiază și controlează noile tehnologii, în special în contextul modificărilor tehnologice recente din AI.
- 00:10:00 - 00:15:00
El discută despre optimismul tehnologic și prezumția economică conform căreia tehnologiile noi, deși disruptivante, aduc beneficii generale. Introduce conceptul de „productivity bandwagon”, care sugerează că noile tehnologii ar trebui să ducă la creșteri salariale reale.
- 00:15:00 - 00:20:00
Darren argumentează că această presupunere este simplificată și adesea ignoră aspecte precum cine controlează tehnologia și cum afectează distribuția veniturilor. Folosește exemple istorice pentru a ilustra că productivitatea crescută nu duce întotdeauna la creșteri ale cererii de muncă.
- 00:20:00 - 00:25:00
Indian exemplifică două tehnologii transformative: morile de vânt și ginul de bumbac al lui Eli Whitney. Deși au revoluționat economiile lor, muncitorii nu au beneficiat echitabil, mai ales din cauza forței coercitive și a relațiilor de putere.
- 00:25:00 - 00:30:00
Revoluția industrială britanică este analizată pentru a demonstra că creșterea productivității nu a beneficiat muncitorii în primele faze, iar coerciția și automatizarea au jucat un rol important în acest sens.
- 00:30:00 - 00:35:00
Analiza continuă cu dezvoltarea tehnologică modernă și impactul asupra salariilor reale în SUA, subliniind distrugerea partajării prosperității observată după 1980 și creșterea inegalității în multe economii industrializate.
- 00:35:00 - 00:40:00
Se folosesc exemple din industria auto, unde inițial inovațiile tehnologice au creat locuri de muncă și partajarea prosperității prin sindicate, dar ulterior direcția s-a schimbat spre automatizare, reducând cererea de muncă consolidată cu puterea muncii.
- 00:40:00 - 00:45:00
Discuția se îndreaptă spre influența gândirii ideologice și a priorităților afacerilor, menționând influența lui Milton Friedman și schimbarea accentului către reducerea costurilor și profit pentru acționari, în detrimentul împărțirii câștigurilor.
- 00:45:00 - 00:50:00
Subliniază că beneficiarii tehnologiei sunt determinate de modul în care alegem să dezvoltăm și să utilizăm tehnologiile, având în vedere relațiile de putere și dacă inovațiile conduc la automatizare sau la sarcini noi pentru muncitori.
- 00:50:00 - 00:55:00
Prezintă diferite perspective tehnologice în istoria computerelor, de la vizunea lui Alan Turing către utilitatea tehnologiei pentru oameni discutată de Norbert Wiener, reflectând asupra impactului AI, fie ca un instrument de control, fie ca o oportunitate de îmbunătățire a capacităților umane.
- 00:55:00 - 01:00:00
Disputează polaritatea dintre automare și control, subliniind importanța deciziilor democratice și ale mișcărilor societății civile în determinarea direcției în care se dezvoltă tehnologiile digitale.
- 01:00:00 - 01:05:00
În arrațiunea despre energia regenerabilă demonstrează cum presiunea societală și reglementările au avut un impact asupra reducerii costurilor tehnologiilor verzi, sugerând cum ar putea avea loc un impact similar în AI.
- 01:05:00 - 01:13:23
În final, discuțiile de Q&A ating teme precum concentrarea pieței, impactul AI asupra țărilor cu acces limitat la tehnologie și potențialul de redistribuire a beneficiilor economice, întărind ideea că alegerea noastră colectivă va defini viitorul.
Mind Map
Frequently Asked Question
Cine este Darren Acemoglu?
Darren Acemoglu este profesor la MIT și este renumit pentru cercetările sale în domeniul economiei.
Care este subiectul principal al prelegerii?
Subiectul principal este impactul tehnologiei asupra economiei și cum poate influența distribuția veniturilor.
Ce teme abordează profesorul în prelegere?
Profesorul discută despre relațiile de putere, automatizare și modul în care tehnologia afectează piața muncii.
Există exemple istorice discutate în prelegere?
Da, sunt discutate Revoluția Industrială Britanică și alte inovații tehnologice semnificative din istorie.
Ce poziție are profesorul față de automatizare?
Acemoglu este precaut față de automatizare, subliniind importanța analizării cum și cine controlează noile tehnologii.
Cum afectează tehnologia distribuția veniturilor, potrivit profesorului?
Noul val de tehnologii poate crește inegalitatea veniturilor dacă relațiile de putere nu sunt corecte.
Ce soluții propune pentru o distribuție echitabilă a prosperității tehnologice?
Propune schimbări instituționale și democratice pentru a asigura ca beneficiile tehnologiei să fie distribuite mai echitabil.
Cum vede rolul AI în economie?
AI poate fi un instrument de control și automatizare, dar are și potențialul de a îmbunătăți productivitatea umană dacă este utilizată corect.
Există o soluție pentru adaptarea la noile tehnologii?
Este necesară o implicare mai puternică a societății civile și reglementări guvernamentale eficiente.
Ce exemple moderne de impact al tehnologiei sunt discutate?
Se discută despre impactul AI și al altor tehnologii digitale asupra locurilor de muncă contemporane.
View more video summaries
ZOHO Case Study | बिना किसी लोन के बना डाला भारत का सबसे Profitable Startup | Rahul Malodia
Are you a Nurse Entrepreneur or Intrepreneur
Old & Odd: Archaea, Bacteria & Protists - CrashCourse Biology #35
Empowering a Data Driven Culture: Self Service, Data Literacy and External Analytics
Advocacy Update
Supernatural - Horror Game Playthrough W/ Lui (Dude I'm Not Scared) (Reupload)
- 00:00:08everybody Welcome to uh the Sanaya La
- 00:00:11Memorial lecture with Professor Darren
- 00:00:13asoglu um I'm Professor Andrew Steven
- 00:00:16I'm the deputy Dean for faculty research
- 00:00:18here at the Sai business school and it's
- 00:00:21an absolute pleasure to be uh hosting uh
- 00:00:24Darren uh in his time um as the uh
- 00:00:28visiting Professor the San well visiting
- 00:00:30Professor uh here at Oxford this uh this
- 00:00:33month um so I'm not going to take up too
- 00:00:35much time because we want to hear from
- 00:00:37Darren and then we'll have a Q&A uh with
- 00:00:39the audience um but as I'm sure you're
- 00:00:43all well aware Darren is a uh a very
- 00:00:46very very highly accomplished
- 00:00:48world-renowned um academic in the field
- 00:00:51of Economics uh he is going to share
- 00:00:55some work from his book uh with us this
- 00:00:58evening uh and then we'll have a discuss
- 00:01:00around that Darren is the Institute
- 00:01:02professor at the department of Economics
- 00:01:03at MIT and the sanay visiting Professor
- 00:01:06here in Oxford um and we're just as I
- 00:01:09said delighted to have him here just a
- 00:01:11little bit of background for those of
- 00:01:12you who aren't familiar with uh this
- 00:01:15visiting professorship um which has been
- 00:01:17running for um about a decade um the
- 00:01:20sanal visiting professorship um here in
- 00:01:23oxid was created to honor the memory and
- 00:01:25academic Legacy of Professor Sanel this
- 00:01:28distinguished scholarship ship scheme
- 00:01:30has run for actually as I said over a
- 00:01:32decade and it's invited some of the most
- 00:01:34prestigious International economists
- 00:01:36here to Oxford um with with Darren being
- 00:01:39the most recent um Honore uh under this
- 00:01:43scheme um for those of you who uh don't
- 00:01:46know U much about Professor Zan Al just
- 00:01:48a little a bit about him and his his
- 00:01:51career here as a development Economist
- 00:01:53and professor of Economics here at the
- 00:01:55University of Oxford uh his research
- 00:01:57interests included the impact of foreign
- 00:01:59direct invest ment in developing
- 00:02:01countries the E the economics of
- 00:02:02multinational corporations and the
- 00:02:04development of technological capability
- 00:02:07and Industrial competitiveness in
- 00:02:08developing countries he's one of the
- 00:02:10world's preeminent development
- 00:02:12economists L was also one of the
- 00:02:14founding editors of the Oxford
- 00:02:16development studies journal and also a
- 00:02:19senior Economist at the World Bank and
- 00:02:20so it's a real pleasure that we can
- 00:02:22honor his legacy with this visiting
- 00:02:25professorship um that uh we're we're
- 00:02:28honoring tonight with uh Professor
- 00:02:30Darren Asam MoGo so without further Ado
- 00:02:32I'm going to ask Darren to come on up to
- 00:02:34the elcon and um give us our talk you
- 00:02:37saw the QR code there on the screen um
- 00:02:40actually if we can just quickly bring
- 00:02:42that back um in the Q&A part this
- 00:02:46evening we'll have mics roaming around
- 00:02:48in the lecture theater but if you're
- 00:02:50watching online on the live stream or
- 00:02:53you'd rather not hold a microphone and
- 00:02:55speak but just put something uh online
- 00:02:57then use that uh that QR code which will
- 00:03:00take you to a web page where you can
- 00:03:01submit your questions and then I'll
- 00:03:03moderate the discussion uh with Darren
- 00:03:05and see the questions on an iPad uh when
- 00:03:08the time comes all right Darren over to
- 00:03:11you thank you thank you very much Andrew
- 00:03:13and uh uh it's my true pleasure to be
- 00:03:15here thank you to the S business school
- 00:03:18for hosting me and thank you to the L
- 00:03:20family for uh their great Hospitality as
- 00:03:22well and thank you to all of you for
- 00:03:24being here and it's my pleasure to be at
- 00:03:27Oxford it's also my pleasure to share
- 00:03:29share uh some thoughts from my recent
- 00:03:32book with Simon Johnson power and PR
- 00:03:34progress our Thousand-Year struggle over
- 00:03:37technology and prosperity I think the
- 00:03:39title essentially frames the question
- 00:03:41that I want to talk about today which is
- 00:03:43that we are in the middle of some pretty
- 00:03:47uh impressive changes in
- 00:03:51technology from biotech to communication
- 00:03:55Technologies and especially most
- 00:03:56recently to breakthroughs in AI there is
- 00:04:00tremendous amount of
- 00:04:03excitement
- 00:04:06but often under
- 00:04:10appreciated is how Society adapts and
- 00:04:15works with new technologies in
- 00:04:17particular in regards to the question of
- 00:04:19who will benefit from new technologies
- 00:04:21who will control these new technologies
- 00:04:23the emphasis on
- 00:04:26struggle is for a reason because we
- 00:04:31think this is one of the most important
- 00:04:34axes of understanding the impact of
- 00:04:36technology over society and economy
- 00:04:38especially during this transformative
- 00:04:40age and it's not receiving its due
- 00:04:44attention why are we perhaps
- 00:04:49not worrying as much about who controls
- 00:04:52technology and what technology does
- 00:04:54especially in terms of income
- 00:04:56distribution winners and losers and I
- 00:04:58think there are two sorts of reasons
- 00:05:02especially looked at it from the other
- 00:05:05side of the Atlantic where you see
- 00:05:07techno optimism in the media in policy
- 00:05:10circles has been very very strong that
- 00:05:12somehow we will work out the rough edges
- 00:05:15of the adjustment process to new
- 00:05:17technology and the second is actually on
- 00:05:20both sides of the Atlantic is related to
- 00:05:22the economic science that there is a
- 00:05:25widespread belief in economics that
- 00:05:29while there are disruptive effects of
- 00:05:32Technologies there is a very
- 00:05:36powerful
- 00:05:37mechanism at work which will ultimately
- 00:05:41make sure that we will generally tend to
- 00:05:45benefit all of us will generally tend to
- 00:05:47benefit from new
- 00:05:49technologies and of course since most of
- 00:05:52us actually earn our livings in the
- 00:05:54labor market that mechanism has to go
- 00:05:56through the labor
- 00:05:58market s and I call it the productivity
- 00:06:00bandwagon it is a very very widespread
- 00:06:03idea in
- 00:06:05economics perhaps so widespread that it
- 00:06:07doesn't have a name so we had to
- 00:06:08Christen it what does the productivity
- 00:06:10bone bandwagon says essentially the uh
- 00:06:13argument is very
- 00:06:14simple at least superficially and I'm
- 00:06:17going to argue that it actually gains
- 00:06:20its Simplicity from perhaps ignoring
- 00:06:23some important considerations product
- 00:06:25technology improves so our knowledge
- 00:06:27improves so for example we didn't know
- 00:06:28how to do
- 00:06:30uh large language models before now we
- 00:06:32do if we use that the right way that's
- 00:06:35an that's an if there are Technologies
- 00:06:37in the past we haven't used it for
- 00:06:39improving productivity but if we use it
- 00:06:41the not the way that we use nuclear
- 00:06:43weapons or not the way that we used uh
- 00:06:45other things which uh have been deployed
- 00:06:48for destructive reasons and in the case
- 00:06:50of AI I think there are some questions
- 00:06:52but it's not a crazy presumption then
- 00:06:54that's going to improve
- 00:06:56productivity meaning we're going to be
- 00:06:58able to produce more goods and services
- 00:07:00or higher quality goods and services
- 00:07:03with the same amount of inputs in
- 00:07:04particular with the same amount of Labor
- 00:07:07and then the presumption of the
- 00:07:08productivity bandwagon is
- 00:07:11that that will somehow translate to
- 00:07:15higher real
- 00:07:17wages meaning that now workers are going
- 00:07:19to earn more and are going to have the
- 00:07:23capacity to purchase more goods and
- 00:07:25services so we've now spread all these
- 00:07:28gains more widely in society so why
- 00:07:30would that be the case so if I write it
- 00:07:32this way as I've done here on the left
- 00:07:35it seems like reasonable but you know if
- 00:07:38we dig into it why would it be the
- 00:07:40case so the reasoning
- 00:07:43goes something like this and it's a
- 00:07:45little more subtle than perhaps that
- 00:07:48chart on the left might first
- 00:07:50suggest the idea is that productivity
- 00:07:54increases that makes employers want to
- 00:07:58hire more labor because now labor has
- 00:08:01become more productive the given amount
- 00:08:02of Labor can produce more goods and
- 00:08:05services and if employers rush out
- 00:08:10to demand more
- 00:08:13labor the market process the competitive
- 00:08:16process is going to bid up
- 00:08:18wages now if once you start spelling it
- 00:08:21out this way you'll realize that there
- 00:08:24are two presumptions
- 00:08:27here one
- 00:08:29which I'm going to spend quite a bit of
- 00:08:30time on is that somehow an increase in
- 00:08:34productivity makes workers sorry makes
- 00:08:37firms want to hire more labor I'm going
- 00:08:40to actually question that I'm going to
- 00:08:42question that because it's actually a
- 00:08:44fairly non-trivial step and we don't
- 00:08:47want to be lulled into a presumption
- 00:08:49that's not really true and sounds true
- 00:08:52so we're we're going to dig dig into
- 00:08:54that but I'm going to start from the
- 00:08:56second presumption here on this slide
- 00:08:58and then come to that that one in a
- 00:09:00second which actually is a very
- 00:09:02important
- 00:09:03one that the market process works
- 00:09:06without regard
- 00:09:09to the topics that other social sciences
- 00:09:12worry a lot about for example power who
- 00:09:14has power who has coercive capacity who
- 00:09:16has the ability to impose things on them
- 00:09:19so it may it is a non-trivial
- 00:09:24step that if employers want more labor
- 00:09:27that will immediately autom atically
- 00:09:30without any barriers translate into
- 00:09:32higher real wages and therefore the
- 00:09:34shared Prosperity step will be completed
- 00:09:36well my argument is going to be
- 00:09:38throughout this talk that throughout
- 00:09:41history and this history is really
- 00:09:43important here I'll come back for the
- 00:09:44reasons the reasons why throughout
- 00:09:46history there are instances where you
- 00:09:48see the productivity bandwagon at work
- 00:09:50so it's not a crazy presumption but
- 00:09:53there are also instances where it hasn't
- 00:09:56worked and it hasn't worked for exact
- 00:09:59these two steps that I have just
- 00:10:01outlined breaking down either because
- 00:10:05power related
- 00:10:07considerations have been important
- 00:10:13right
- 00:10:15or because productivity increases don't
- 00:10:17necessarily translate employers wanting
- 00:10:19more labor so let me start with the
- 00:10:21latter so here is an illustration of two
- 00:10:25very transformative Technologies okay
- 00:10:26I'm not going to be in the business of
- 00:10:28uh you you know what Silicon Valley Tech
- 00:10:30leaders do it's like AI is more
- 00:10:32important than fire I'm not going to say
- 00:10:34AI is more important or less important
- 00:10:35but these are pretty important new
- 00:10:36technologies the one on the left is the
- 00:10:40perhaps the most transformative
- 00:10:42breakthrough of the Middle Ages where
- 00:10:45you know the Dark Ages so-called because
- 00:10:47of cultural reasons were not dark
- 00:10:49technologically there were many many
- 00:10:50many Innovations in the production
- 00:10:52process that really transformed
- 00:10:53agricultural production but none was as
- 00:10:55important as this thing which is
- 00:10:57windmills that really tremendously
- 00:11:00increased the ability of people to use
- 00:11:02energy for a variety of
- 00:11:05tasks the right one the the one on the
- 00:11:08right hand side is equally important is
- 00:11:10Eli Whitney's cotton gin which enable
- 00:11:12the type of cotton that could be grown
- 00:11:14in the US South to be cleaned both of
- 00:11:17these Technologies were truly
- 00:11:19revolutionary in terms of their economic
- 00:11:21consequences they really changed the
- 00:11:22organization of production in many parts
- 00:11:24the windmill in much of medieval Europe
- 00:11:28the uh uh Eli Whitney's cut engin or
- 00:11:31actually Eli Whitney was one of the
- 00:11:32people who invented the cut engine there
- 00:11:34were several others but Eli Whitney gets
- 00:11:36the credit uh the the cut engine
- 00:11:39completely transformed the US southern
- 00:11:41economy it was a complete Backwater
- 00:11:44economically it became the largest
- 00:11:45exporter of cotton in the world and uh
- 00:11:49engine of the British Industrial
- 00:11:51Revolution which was mostly about cotton
- 00:11:53textile in the early phases in both
- 00:11:57cases and you want f much of a
- 00:11:59productivity bandwagon the medieval
- 00:12:01period despite the series of Innovations
- 00:12:03including the windmill did not see much
- 00:12:06improvements in the real living
- 00:12:07standards of the workers who were the
- 00:12:10farmers and if you think about that for
- 00:12:13a few minutes you'll realize why that's
- 00:12:17quite expected actually because the much
- 00:12:20of the medieval economy was in a
- 00:12:23coercive environment workers were often
- 00:12:25in survi relations or even if when they
- 00:12:28were not in survi Rel
- 00:12:30they had a lot of obligations to uh uh
- 00:12:35uh uh big land owners and and and other
- 00:12:38powerful actors such as the uh uh the
- 00:12:41church
- 00:12:42hierarchy and when more labor was
- 00:12:47demanded those obligations could be
- 00:12:49increased via the coercive power of the
- 00:12:53high aristocracy and the church
- 00:12:55hierarchy and the evidence seems to
- 00:12:58suggest that activity increased a lot
- 00:13:00but a lot of it was captured by a very
- 00:13:02small fraction of the population about
- 00:13:045% or so who went around building
- 00:13:06wonderful monuments such as the ones you
- 00:13:08see in Oxford but people remained
- 00:13:11poor the one on the right is even
- 00:13:13clearer to see who were the workers they
- 00:13:16were the black enslaved workers and of
- 00:13:20course when sudden plantation owners
- 00:13:24decided oh my God now there is a great
- 00:13:26Economic Opportunity in the form of
- 00:13:28cotton Plantation they didn't tell the
- 00:13:31slaves oh well how much more money
- 00:13:32should we give you they told them okay
- 00:13:35fine you're moving to the Mississippi
- 00:13:36Delta the down deep south where
- 00:13:38conditions were much worse long hours uh
- 00:13:41much more coercive relations and uh and
- 00:13:45living standards of the slaves actually
- 00:13:47worsened now you might think these two
- 00:13:51examples
- 00:13:53are
- 00:13:56non-representative because precisely
- 00:13:59they are non-industrial Technologies
- 00:14:01although the cotton genin is sort of
- 00:14:03semi-industrial or it's adjacent to
- 00:14:06Industrial and precisely because I have
- 00:14:10emphasized coercion but coercion is just
- 00:14:13the tip of a broader Iceberg which is
- 00:14:15power power relations are as important
- 00:14:18in every relation every production
- 00:14:20relation not just in the medieval
- 00:14:22economy but perhaps even more
- 00:14:25quintessential for understanding
- 00:14:27technology is the Industrial Revolution
- 00:14:29ution after all the reason why Simon
- 00:14:33Johnson and I put so much
- 00:14:37emphasis in the book on history is
- 00:14:42because my experience I think Simon's
- 00:14:44experience also is that for the last 15
- 00:14:47years when I talked about robots or
- 00:14:50digital Technologies or other things
- 00:14:53creating huge inequality wage declines
- 00:14:56uh and and being misused
- 00:14:59you would always get especially from
- 00:15:01journalists in the US and tech people oh
- 00:15:04you must be saying this time is
- 00:15:05different because we know in history
- 00:15:07everything has worked out look at the
- 00:15:08British Industrial Revolution well I
- 00:15:10don't know what British Industrial
- 00:15:12Revolution they were looking at because
- 00:15:13the history is actually very different
- 00:15:15the history of the British Industrial
- 00:15:16Revolution
- 00:15:17is
- 00:15:19not devoid of
- 00:15:22coercion the factory system was a
- 00:15:24coercive system at least in the early
- 00:15:27phases but
- 00:15:29most importantly it shows great
- 00:15:32parallels to the other two examples that
- 00:15:34I showed earlier on the British industri
- 00:15:36revolution for all practical purposes
- 00:15:38started sometime around 1740 1750 when
- 00:15:43Advanced Machinery started being applied
- 00:15:45developed and applied in uh in the
- 00:15:46production process and the evidence is
- 00:15:49that by and large we don't see much
- 00:15:51improvement in the real earnings
- 00:15:54inflation adjusted price adjusted
- 00:15:56earnings of the working class for about
- 00:15:589 years up to about 1840 or so and the
- 00:16:02most dynamic sectors of the economy for
- 00:16:04example weaving that text the Cottons
- 00:16:08that were being or uh cotton yarn that
- 00:16:11was being exported or cotton that was
- 00:16:14being exported and turned into yarn by
- 00:16:15Spinners and then Weavers were
- 00:16:17processing that they actually during the
- 00:16:20process of the early industrialization
- 00:16:22saw their real earnings decline by about
- 00:16:26half moreover the
- 00:16:31factory was indeed partly about control
- 00:16:35and as a result the factory was a very
- 00:16:37oppressive place and the working hours
- 00:16:40of typical British worker increased by
- 00:16:43about 20% so on an hourly basis real
- 00:16:48wages probably declin for about a good
- 00:16:5090 years so the sort of productivity
- 00:16:54bandwagon saving us all you know for AI
- 00:16:57just wait for another two years well if
- 00:16:59the British Industrial Revolution is the
- 00:17:01example we have to learn from doesn't
- 00:17:03look so
- 00:17:04good so what happened in the British
- 00:17:07Industrial Revolution well again I
- 00:17:09mentioned it's the power issue was very
- 00:17:12important that's why it's in the title
- 00:17:13of our book but there's another aspect
- 00:17:16British industrial revolution's early
- 00:17:19Technologies were all centered on
- 00:17:22automation meaning substitution of
- 00:17:25machinery for tasks previously performed
- 00:17:28by workers
- 00:17:31and what's so interesting about
- 00:17:32automation is that
- 00:17:35a this is a lot of what we're doing with
- 00:17:38digital Technologies today and at least
- 00:17:40one path of what we might be doing with
- 00:17:44AI and
- 00:17:46B it actually if you think about it
- 00:17:49starts
- 00:17:50questioning that other part of the
- 00:17:53causal chain that productivity Rises and
- 00:17:56then as productivity Rises employers
- 00:17:58would like to hire more labor in fact
- 00:18:02the reason why hen
- 00:18:04Weavers did so well in terms of their
- 00:18:07they were the labor aristocracy because
- 00:18:08there was great demand for weaving
- 00:18:11skills and why did their real incomes
- 00:18:14decline
- 00:18:15by by by declined by half well because
- 00:18:19once weaving was mechanized
- 00:18:21automated their skills were not
- 00:18:24necessary so actually the big step here
- 00:18:28doesn't follow so if I wanted to put it
- 00:18:31in a slightly more wonkish terms when
- 00:18:33people say productivity Rises and all
- 00:18:36economists well many economists and all
- 00:18:39journalists are guilty of this because
- 00:18:41they're actually confusing two things
- 00:18:44they're confusing average productivity
- 00:18:46which is how much output is produced per
- 00:18:48worker versus what economic theory tells
- 00:18:51us what firm's value which is the
- 00:18:54marginal productivity of workers meaning
- 00:18:56what a worker contributes and why
- 00:18:59these two Notions could go very
- 00:19:01different ways well I think the best way
- 00:19:03of understanding that is this is this
- 00:19:05story which is supposed to be a humorous
- 00:19:07story well unless you're a worker of
- 00:19:08course uh which is
- 00:19:11the it says the factory of the future
- 00:19:14will have two employees a man and a dog
- 00:19:17the man is there to feed the dog and the
- 00:19:18dog is there to make sure that the man
- 00:19:20doesn't touch the
- 00:19:21equipment
- 00:19:23so if indeed that is the future of the
- 00:19:27factory well it's not very humorous for
- 00:19:29the workers but it will have huge
- 00:19:31average labor productivity you'll have
- 00:19:33one worker or perhaps two if you count
- 00:19:35the dog and a huge amount of output this
- 00:19:37amazing Machinery is producing a
- 00:19:40lot but the reason why this a humorous
- 00:19:42story is that if the Machinery gets
- 00:19:44better and you double output of course
- 00:19:46firms are not going to rush out to hire
- 00:19:47more men and their dogs the marginal
- 00:19:50productivity the contribution of the
- 00:19:52worker is
- 00:19:53Trivial so if
- 00:19:55automation is the order of the day
- 00:19:59it takes tasks away from workers to give
- 00:20:02it to Capital the reason why weaving
- 00:20:04Machinery was so useful is because
- 00:20:06Weavers were expensive they were the
- 00:20:07labor
- 00:20:08aristocracy you don't need the Weavers
- 00:20:10anymore now Machinery does it that's
- 00:20:13productive you produce more output but
- 00:20:16it's just like the guy and his dog
- 00:20:18they're not
- 00:20:19needed so there are therefore two big
- 00:20:24steps here that we have to think about
- 00:20:26the productivity bandwagon in the past
- 00:20:29and in the future in the age of
- 00:20:32AI if we are going to hope that somehow
- 00:20:35the market process the automatic process
- 00:20:37is going to get us shared Prosperity it
- 00:20:40must be through the productivity
- 00:20:42badwagon
- 00:20:43or some institutional responses I'll
- 00:20:45come back to them and if it's going to
- 00:20:46be the productivity bandwagon then we
- 00:20:48need those two steps to work out somehow
- 00:20:50new technologies will increase the
- 00:20:52marginal productivity of Labor and power
- 00:20:54issues are not going to make Capital
- 00:20:56dominant over labor
- 00:20:59these issues were of course Very Much
- 00:21:02alive and very much understood during
- 00:21:03the British Industrial Revolution Jeremy
- 00:21:06benam came up with his
- 00:21:08panopticon precisely arguing that this
- 00:21:11was an efficiency enhancing technology
- 00:21:13but many people understood right away
- 00:21:16that this was actually a technology of
- 00:21:17control it was a way of taking power
- 00:21:21away from
- 00:21:23workers and giving it more to Capital
- 00:21:26and part of the reason why work fared
- 00:21:29badly wasn't just working hours and low
- 00:21:31incomes but because life expectancy fell
- 00:21:33as low as 30 to 30 years at Birth in
- 00:21:35cities like Manchester and London why
- 00:21:38because they were shoved in into cities
- 00:21:40with horrible working conditions and
- 00:21:41nobody gave them public infrastructure
- 00:21:43clean toilets or any sort of uh
- 00:21:46amenities so these were all about
- 00:21:49power power is going to be very
- 00:21:51important in the future but this issue
- 00:21:53of automation is going to be very
- 00:21:54important as
- 00:21:56well but of course when people tell you
- 00:21:59and I'll tell you
- 00:22:00that we are of course today in the year
- 00:22:042024 we are hugely fortunate that the
- 00:22:07British Industrial Revolution happened
- 00:22:09you know uh 200 years ago or start 250
- 00:22:14years ago started uh we are much
- 00:22:17healthier We Have Much Better Health
- 00:22:19Technologies we have much better comfort
- 00:22:21and we have we have much much much much
- 00:22:22greater levels of real incomes so what
- 00:22:25happened well things changed after 184
- 00:22:2840 and how did they change they changed
- 00:22:31because institutions
- 00:22:35changed especially power relations with
- 00:22:37democracy and trade unions in Britain
- 00:22:40trade unions were very heavily
- 00:22:41prosecuted up to the last quarter of the
- 00:22:4319th century and the nature of
- 00:22:46technology changed but let me actually
- 00:22:49stop talking about the British
- 00:22:50Industrial Revolution you could get
- 00:22:52boring so let me talk about the modern
- 00:22:55times let me talk about the modern times
- 00:22:58and make the case that actually all of
- 00:23:02these issues are not just hypothetical
- 00:23:03theoretical things we have to sort of
- 00:23:05think about for historical
- 00:23:07Nuance this is real wages for 10
- 00:23:11demographic groups in the United
- 00:23:14States and it's for men and women men
- 00:23:18here women here and different curves are
- 00:23:21for different education groups workers
- 00:23:23without a high school degree high school
- 00:23:24graduates workers with associate degrees
- 00:23:26just a college degree and those with
- 00:23:27postgraduate degrees everything is
- 00:23:30normalized to zero so you see the
- 00:23:31cumulative real weage changes now these
- 00:23:34figur shows two very contrasting period
- 00:23:37of periods of uh economic growth in the
- 00:23:40west or in the United States the other
- 00:23:42countries are I'll show you a little bit
- 00:23:44on that but this is the end of the
- 00:23:48period of shared Prosperity that had
- 00:23:49started know arguably in the 40s but uh
- 00:23:55certainly you see it very strongly in
- 00:23:56the 1950s but in this data set i'm
- 00:23:58showing you from 1963 onwards you see
- 00:24:00all these curves are actually on top of
- 00:24:02each other you have real wage growth for
- 00:24:05all of 10 of these demographic groups
- 00:24:07are essentially parallel which means
- 00:24:09their real wages are growing at the same
- 00:24:10rate and they're growing very rapidly
- 00:24:12how how rapid it's actually a
- 00:24:13mindboggling 2 and a half% a year in
- 00:24:16real terms that's faster than
- 00:24:18productivity growth so labor was
- 00:24:19benefiting from productivity growth
- 00:24:21during this period as much as perhaps
- 00:24:23slightly more than Capital that was the
- 00:24:26center tenant of shared Prosperity so
- 00:24:29you see the kind of process that in the
- 00:24:3219th century in the early 20th century
- 00:24:35was the basis of that period of shared
- 00:24:38prosperity and you see it's oops and you
- 00:24:40see it's undoing from around 1980 these
- 00:24:43curves are going in different direction
- 00:24:45but more striking is that the real wages
- 00:24:48of many workers actually about half of
- 00:24:51the US Labor Force is actually declining
- 00:24:53as productivity is increasing so this is
- 00:24:56a very jarring version of of non-shared
- 00:25:01growth so therefore it confronted with
- 00:25:04this picture a natural instinct is to
- 00:25:07say why was there this shed prosperity
- 00:25:10and why did it come apart so let me try
- 00:25:12to answer that but before doing that let
- 00:25:15me say that this is not the same across
- 00:25:18countries but it's not unique to the
- 00:25:19United States so in particular pretty
- 00:25:21much every industrialized nation is
- 00:25:23struggling with the same pattern
- 00:25:25inequality has increased in most
- 00:25:26countries us was the leader of
- 00:25:28inequality and Remains the leader of
- 00:25:30inequality but many other countries have
- 00:25:32also
- 00:25:34undergone similar processes and many
- 00:25:37countries have struggled by the same
- 00:25:40struggled with the same problem of
- 00:25:43stagnant or sometimes even declining
- 00:25:45like in Germany or in the UK real
- 00:25:48incomes of people at the bottom of the
- 00:25:50income
- 00:25:51distribution so here let me go from the
- 00:25:55British Industrial Revolution to the
- 00:25:57early 20s Century
- 00:25:59industrialization and go to the iconic
- 00:26:01sector of the car manufacturing because
- 00:26:04that was a leader in IND introducing
- 00:26:07machinery and it was also a leader in
- 00:26:10illustrating the types of shared
- 00:26:14Prosperity dynamics that I emphasized
- 00:26:16already in
- 00:26:17particular you see here a picture from
- 00:26:20the Ford motor Factory from 1919 Henry
- 00:26:23Ford was at the very Forefront of
- 00:26:25introducing mass production which meant
- 00:26:27Machinery so he was doing the automation
- 00:26:29he was introducing the he was you know
- 00:26:32his factories were the first time there
- 00:26:35were the gun Armament producers and
- 00:26:39sewing machine producers that were doing
- 00:26:40the same thing but he did at a much
- 00:26:43bigger scale introducing the
- 00:26:44interchangeable part system so some more
- 00:26:47standardized production he introduced
- 00:26:49decentralized electrical Machinery so
- 00:26:50that machines could work much better and
- 00:26:53much more
- 00:26:53efficiently so he was very much in the
- 00:26:56business of automation but also also
- 00:26:58what you see with Henry Ford's in Henry
- 00:27:01Ford's Factory is that while doing that
- 00:27:04he also introduced and his Engineers
- 00:27:06also introduced a completely different
- 00:27:07way of in doing work much more technical
- 00:27:10work much more engineering knowledge
- 00:27:12both at the back office and on the front
- 00:27:15line so no surprise that when you see
- 00:27:17pictures from H Henry Ford's motor
- 00:27:19Factory you see workers doing new tasks
- 00:27:22new activities where their marginal
- 00:27:25productivity was really Central for the
- 00:27:27production process
- 00:27:29as a result of these
- 00:27:31tasks as a tremendous amount of new
- 00:27:34machinery and investment was introduced
- 00:27:36in the Auto industry employment
- 00:27:38increased by
- 00:27:3910-fold so this was employment growth
- 00:27:42together with lots of new Machinery
- 00:27:44being introduced but the second one was
- 00:27:47very much related to power the Auto
- 00:27:49industry wasn't just a leader in
- 00:27:51Automation and new tasks but it was also
- 00:27:53a leader in World labor organization in
- 00:27:56a series of
- 00:27:58uh iconic events labor organization
- 00:28:01became most powerful in the Auto
- 00:28:03industry and from there it spread to the
- 00:28:06rest of the uh for to the rest of the
- 00:28:08manufacturing industry there were a
- 00:28:10number of very important ones uh this
- 00:28:13one is the uh there was one in uh uh uh
- 00:28:17in in Ford motor factories this is a
- 00:28:19slightly later one you this is called
- 00:28:21the sitdown strike where workers sat
- 00:28:22down rather than doing the work and the
- 00:28:24GM General Motors had to agree with
- 00:28:26better working conditions more technical
- 00:28:28work and and higher
- 00:28:31wages so therefore my argument is that
- 00:28:34the reason why we had this type of
- 00:28:37growth both in the first half of the
- 00:28:3920th century and in the decades
- 00:28:41following World War II is precisely
- 00:28:43because these two pillars of the
- 00:28:44productivity bandwagon not just automate
- 00:28:48work but create work for workers and the
- 00:28:51power relations at least giving some
- 00:28:53voice so that workers could not be
- 00:28:56forced to work at low wages were both
- 00:28:58important so why did then the shared
- 00:29:01Prosperity come apart well because those
- 00:29:04two pillars came apart here is how it
- 00:29:06auto Factory looks like except for the
- 00:29:09color this Machinery is very similar to
- 00:29:12what Ford was doing it's a automation
- 00:29:16Machinery during Ford's days it was
- 00:29:18putting things together in very
- 00:29:20rudimentary way that started being
- 00:29:22mechanized here it's painting welding as
- 00:29:25well as operation and assembly but
- 00:29:27what's remarkable in this Auto Factory
- 00:29:29is that you don't see the workers doing
- 00:29:30the technical tasks so many employers
- 00:29:34introduced the automated
- 00:29:37Machinery but they didn't go to the next
- 00:29:41step of introducing the new tasks hence
- 00:29:43we took one more step towards that
- 00:29:45modern Factory with the man and the
- 00:29:47dog so technology became more biased
- 00:29:51towards labor more substituting for
- 00:29:53labor now of course the type of increase
- 00:29:57inequality is an epocal event has many
- 00:30:00causes but automation is really at the
- 00:30:02heart of it and that's what my work with
- 00:30:04Pascal Restrepo has argued so here is a
- 00:30:07summary of that using this chart where
- 00:30:09what I'm doing is I'm looking at more
- 00:30:12detailed demographic groups distinguish
- 00:30:14not just by gender and education but
- 00:30:16gender education ethnicity and age and
- 00:30:18I'm zeroing on that period from 1980 to
- 00:30:21to the just before the covid period and
- 00:30:24on the vertical axis I'm showing the
- 00:30:26cumulative change in the real wages and
- 00:30:29you can see that many of these circles
- 00:30:30actually very big circles here meaning
- 00:30:32big demographic groups are below the
- 00:30:34zero line those are the people who are
- 00:30:36losing out in terms of real incomes as
- 00:30:39the economy is growing in terms of
- 00:30:40productivity and the new element here is
- 00:30:43the horizontal axis that's my and
- 00:30:46pasqual's estimate of how much
- 00:30:50automation has impacted that demographic
- 00:30:52group it's roughly speaking the fraction
- 00:30:54of tasks that the relevant demographic
- 00:30:56group was performing in 1980 that have
- 00:30:58since been automated and what you see is
- 00:31:00a very strong relationship you know if
- 00:31:02you want to understand why these groups
- 00:31:04are doing so badly I think our evidence
- 00:31:06shows you cannot do that without seeing
- 00:31:08that uh these groups are the ones whose
- 00:31:11jobs have been whose tasks have been
- 00:31:13automated but power is also at the
- 00:31:16center of it and power I want to now
- 00:31:21argue
- 00:31:23is really about two things in the modern
- 00:31:27world it is
- 00:31:29about
- 00:31:31organization so United Auto Workers
- 00:31:33sitdown strike was a step towards
- 00:31:35strengthening labor so that they took a
- 00:31:37slice from the productivity increases
- 00:31:40but this the iconic event of the 1980s
- 00:31:43the labor relation starts with the Petco
- 00:31:45strike per professional air traffic
- 00:31:46controller strike of 1981 where Ronald
- 00:31:49Reagan as the new president fired all
- 00:31:50the professional air traffic controllers
- 00:31:52and that was a signal to many other
- 00:31:54employers who similarly took a very
- 00:31:56harsh line against TR strike activity
- 00:31:59declin labor power decline unionization
- 00:32:01started an accelerated decline after
- 00:32:03that point although it was already
- 00:32:04declining because of the decline the the
- 00:32:07diminishing importance of
- 00:32:08manufacturing but Power when I talk
- 00:32:11about it is not just about unions it's
- 00:32:14not certainly today in the United States
- 00:32:17it's not about coercion you know there's
- 00:32:18no slavery nobody in survi relations not
- 00:32:21not many people anyway but it is also
- 00:32:24about
- 00:32:25ideology what sort of priority powerful
- 00:32:28actors
- 00:32:30have and the two book hens that
- 00:32:35determined how gains in manufacturing
- 00:32:38how gains in Industry got divided was
- 00:32:43this but also the changing priorities
- 00:32:46and ideologies of
- 00:32:49businesses you can call it ideology you
- 00:32:52can call it something else in the book
- 00:32:53we call it Vision because ideology
- 00:32:55sometimes has some bad
- 00:32:56connotations but it was the changing
- 00:33:00emphasis among managers and among owners
- 00:33:05that good corporations were the lean
- 00:33:08corporations that cut labor costs didn't
- 00:33:11share the benefits with workers because
- 00:33:13that would be waste and would reduce the
- 00:33:15what the shareholders got and the more
- 00:33:16you gave to shareholders the more
- 00:33:17Dynamic the corporations became the more
- 00:33:20Investments it led to and so on and so
- 00:33:22forth and of course the symbol of that
- 00:33:24was this great Economist Milton Freedman
- 00:33:26but Milton Friedman's most famous
- 00:33:28article wasn't one of those in scholarly
- 00:33:30journals but it's something he wrote In
- 00:33:32The New York Times magazine where he
- 00:33:34said the only social responsibility of
- 00:33:35business is to look after his
- 00:33:36shareholders cut costs cut Wages that's
- 00:33:39what's good and uh and it was a rallying
- 00:33:42cry for many managers who for a variety
- 00:33:45of reasons partly unions partly other
- 00:33:48considerations Norms felt obliged that
- 00:33:52if they make more profits if they export
- 00:33:53more they'll share some of that with the
- 00:33:55workers well fredman was telling them
- 00:33:57you have license you don't have to do
- 00:33:58that who will stand up against them well
- 00:34:02unions could have done but unions were
- 00:34:03also in Decline so these two book meant
- 00:34:06that the power relations shifted against
- 00:34:10labor so the reason why I went into some
- 00:34:13details on these topics is
- 00:34:16because I think to understand who will
- 00:34:19benefit
- 00:34:20from new technologies it is really
- 00:34:24important to consider these two things
- 00:34:26that have been drivers of who gets the
- 00:34:30benefits in history whether new
- 00:34:33technologies just automates or
- 00:34:35introduces new tasks and hence increases
- 00:34:37the marginal product of labor and how we
- 00:34:39def power relations but when we think
- 00:34:42about Technologies when I'm talking
- 00:34:43about what here technology has two
- 00:34:47parts what businesses are doing and
- 00:34:50that's what I put the emphasis on way by
- 00:34:52saying uh Milton Freedman you know this
- 00:34:54is what you know people like Jack Welch
- 00:34:57who were running these big businesses
- 00:34:59but businesses of course can do with
- 00:35:03technology what the nature of technology
- 00:35:06or what the characteristics of the
- 00:35:08technology allow and that is
- 00:35:10determined by what the tech sector is
- 00:35:13doing and that's becoming even more
- 00:35:15important today with AI and advances in
- 00:35:17digital Technologies and you might think
- 00:35:20if the world is just one of seamlessly
- 00:35:22working
- 00:35:24markets all that that's important is
- 00:35:26what's the profitable
- 00:35:28strategy or perhaps if you believe Sam
- 00:35:30Altman and uh and Mark Zuckerberg they
- 00:35:33are also altruistic
- 00:35:35so but actually in the same way that
- 00:35:39ideology Vision priorities matter if you
- 00:35:41believe fredman you cut wages if you
- 00:35:45believe in what people used to call
- 00:35:47welfare capitalism you share the gains
- 00:35:48with the
- 00:35:49workers well perhaps what the visions of
- 00:35:54what the tech leaders do matters as well
- 00:35:57and in fact in fact the argument in the
- 00:35:58book and the reason for why I am worried
- 00:36:01about the future is that it's not just a
- 00:36:03problem of corporate America it's even
- 00:36:05worse problem of tech America and the
- 00:36:09problem of tech America which decides of
- 00:36:11course the trajectory of technology is
- 00:36:13not just for the us but throughout the
- 00:36:15world is that as
- 00:36:18a call it ideology call it Vision
- 00:36:21preoccupation the whole tech industry
- 00:36:24has become completely focused on
- 00:36:26automation
- 00:36:27why well I don't think you can
- 00:36:29understand that without going back to
- 00:36:30the origin stories of computer science
- 00:36:32and AI it was This brilliant
- 00:36:35mathematician Alan Turing who founded in
- 00:36:39many ways the field of computer science
- 00:36:41but he also had a very different way
- 00:36:44according to you know relative to people
- 00:36:46to his contemporaries of thinking about
- 00:36:48both the human mind and the machines he
- 00:36:50said his work was on computation he said
- 00:36:53the human mind is just a Computing
- 00:36:54machine and computers are Computing
- 00:36:57machines we're going to build better and
- 00:36:58better computers and ultimately the
- 00:36:59computers are going to be as good as the
- 00:37:01computer Computing machine that's in
- 00:37:02your brain so that
- 00:37:05framed the field of computer science as
- 00:37:08one in which you judge the quality of
- 00:37:12the machines by how much human parity
- 00:37:15they achieve meaning how close they
- 00:37:17become to humans and of course if you do
- 00:37:19that it creates a natural bias for
- 00:37:22automation the more human like it is or
- 00:37:25the more it can take over human tasks
- 00:37:29the better and that
- 00:37:33became also the founding vision of the
- 00:37:36field of AI where the the field was
- 00:37:39defined created in
- 00:37:411956 uh before then nobody talks there's
- 00:37:43no term of artificial intelligence these
- 00:37:45gentlemen here who are also very famous
- 00:37:48Define that field and they decare their
- 00:37:50aim to be we're going to reach to human
- 00:37:52brain and human capabilities in six
- 00:37:54months okay perhaps a year all right for
- 00:37:57a little bit optimistic but this vision
- 00:38:00is completely dominant today in the tech
- 00:38:03World artificial general intelligence
- 00:38:06being its most
- 00:38:08recent Offspring but even artificial
- 00:38:10before artificial general intelligence a
- 00:38:12lot of software systems a lot of
- 00:38:15production processes that were digitized
- 00:38:18were all about can we get computers to
- 00:38:21do more of the tasks of humans and that
- 00:38:23creates this bias towards Automation and
- 00:38:24if the tech sector has the bias from
- 00:38:26automation what will corporate leaders
- 00:38:28do so there is a nuts and bolts problem
- 00:38:32if you want to build a
- 00:38:35better tool you need both nuts and bolts
- 00:38:37but if there aren't the nuts aren't
- 00:38:39there nobody is producing the uh
- 00:38:42technologies that are going to be useful
- 00:38:44for
- 00:38:45humans bolts are useless but if bolts
- 00:38:48are not there nobody wants to generate
- 00:38:49the knots either so if if the corporate
- 00:38:51leaders are believed to be just in the
- 00:38:53business of cutting costs now if I stop
- 00:38:57the at this
- 00:38:59slide it would be very Bleak I'm saying
- 00:39:03oh well you know ideology plus business
- 00:39:05models they are condemning us to more
- 00:39:07and more Automation and more and more
- 00:39:08lower wages well
- 00:39:10actually but from the very
- 00:39:14beginning there was a very different
- 00:39:16view of what technologies are about it
- 00:39:19was as early as Alan Touring that
- 00:39:22Norbert weiner a uh another brilliant
- 00:39:24mathematician now this time at MIT was
- 00:39:26writing about humans and computers
- 00:39:29working together and in the book Simon
- 00:39:32Johnson and I call this machine
- 00:39:34usefulness to contrasted with the
- 00:39:36machine intelligence and if you look at
- 00:39:38history history of computers many of the
- 00:39:41things that we completely depend on
- 00:39:42today the computer mouse out of Douglas
- 00:39:44angle Bart uh hyperlink hypertext menu
- 00:39:47driven computers the arpanet and the
- 00:39:49internet uh internet jcr lick liers work
- 00:39:53they all came out to a very different
- 00:39:54Vision where machines were not supposed
- 00:39:56to replace humans
- 00:39:57there wasn't uh human parity Obsession
- 00:40:01but it was machines to be useful to
- 00:40:04humans so the reason why I'm saying this
- 00:40:07is because AI actually increases the
- 00:40:11capabilities both for
- 00:40:13automation if you want to try to
- 00:40:16replicate
- 00:40:18humans large language models are your
- 00:40:20tool but it also increases amplifies the
- 00:40:24possibilities for providing better
- 00:40:26information to humans so so that they
- 00:40:27can comp perform more complex tasks they
- 00:40:29can make better decisions so therefore
- 00:40:32this is actually a contrast between what
- 00:40:36the tech industry is going to be about
- 00:40:38that will matter greatly and choice is
- 00:40:42going to be all over it now the problem
- 00:40:45is that it's actually even worse if you
- 00:40:48get to hang up with uh uh with machine
- 00:40:53intelligence because what it we will end
- 00:40:55up doing is that because you're so
- 00:40:58convinced that machines are better than
- 00:41:00humans you're going to rush to automate
- 00:41:02a lot of tasks and what the US evidence
- 00:41:05shows is that in almost every wave of
- 00:41:09Technologies automation disappointed in
- 00:41:12terms of
- 00:41:12productivity so every wave of
- 00:41:17Technology main frames Innovation uh
- 00:41:20inventory
- 00:41:21systems uh personal computers uh
- 00:41:24software systems for office work they
- 00:41:27all were introduced with great funfare
- 00:41:29productivity is going to double and in
- 00:41:32all cases it was lackluster productivity
- 00:41:35and the reason is quite obvious if
- 00:41:37humans are not as bad as you think and
- 00:41:38machine intelligence is not so great
- 00:41:41you're going to be over automating or
- 00:41:43you're going to be excessively
- 00:41:44automating but the problem is actually
- 00:41:46worse also because again I keep coming
- 00:41:50back the factory system was automating
- 00:41:52but it was also about control but what
- 00:41:54is ai ai is a control machine is AI is
- 00:41:59an information tool and every
- 00:42:00information tool just like Jeremy
- 00:42:03bentham's panopticon is also a tool for
- 00:42:05control so these two may look a world
- 00:42:09apart the social credit system for
- 00:42:12trains in
- 00:42:13China and Facebook so some people will
- 00:42:17say you know social media
- 00:42:19democratization versus centralized
- 00:42:21control but actually there's more
- 00:42:22parallel between these two than meets
- 00:42:24the eye both of those are that one or
- 00:42:27organization with its own ideology is
- 00:42:29centralizing information and it's
- 00:42:31deciding what you see here it's telling
- 00:42:35you well you see only the posts that are
- 00:42:38approved and we're going to incentivize
- 00:42:40that by giving you a higher credit if
- 00:42:42you give the right posts and here you
- 00:42:43see what we decide you see for purposes
- 00:42:46of monetizing that with digital ads or
- 00:42:48whatever El values the people who are
- 00:42:50doing the content moderation in Facebook
- 00:42:52or Google have so power which said was
- 00:42:57Central is also completely Inseparable
- 00:43:00from how we use AI so we're going to
- 00:43:03have these choices who is going to get
- 00:43:05more power with the use of AI and
- 00:43:08whether we're going to use AI to
- 00:43:09increase the marginal productivity and
- 00:43:11the capabilities of workers or we're
- 00:43:12going to try to sideline the workers and
- 00:43:15history shows
- 00:43:17we've have a lot of choices and today we
- 00:43:20have a lot of choices for every example
- 00:43:23or okay fine every 10 example of
- 00:43:25automation I can find one good example
- 00:43:27where some company or some sector has
- 00:43:30used the same Technologies for creating
- 00:43:32new tasks for workers for every Facebook
- 00:43:35Google Chinese Communist party there is
- 00:43:37the system in Taiwan where they have
- 00:43:39used AI in order to make technology more
- 00:43:43pro-democratic so there are choices what
- 00:43:47choices are we going to
- 00:43:48make
- 00:43:50well that depends on who's going to make
- 00:43:52the choices I think if the choices are
- 00:43:53between Sam Alman and E musk I don't
- 00:43:56think we're going to get the right
- 00:43:58outcomes so the history again this time
- 00:44:02is no different history teaches us one
- 00:44:04thing Democratic process is really
- 00:44:07important so if the debate is just
- 00:44:09between Elon Musk and Sam Alman they can
- 00:44:11go in the ring if they want but that's
- 00:44:13not going to bring democracy and if you
- 00:44:15don't bring democracy I don't think
- 00:44:17you're going to have the type of broad
- 00:44:19set of voices that say well we should
- 00:44:22structure this technology both in terms
- 00:44:24of the production process and in terms
- 00:44:26of the relations of power such that a
- 00:44:28broadc cross-section of society
- 00:44:30including the developing World by the
- 00:44:31way who that always gets invol that's
- 00:44:34ignored that these people will benefit
- 00:44:36so we need a democratic process where
- 00:44:39will that Democratic process come from
- 00:44:41well again labor movements but you know
- 00:44:43I'm not sure whether Amazon and
- 00:44:45Starbucks unionization is going to be
- 00:44:47the kind of model so I think we need a
- 00:44:50different approach to labor movement and
- 00:44:52labor voice actually there is some
- 00:44:54interesting changes in the United States
- 00:44:56taking place where where union leaders
- 00:44:58over the last two years have become much
- 00:44:59more concerned about can we make the
- 00:45:02labor movement be an input into the
- 00:45:04technology Direction so that's the
- 00:45:07thesis that I am defending but it will
- 00:45:10also come
- 00:45:11from Civil Society activism so Ral nater
- 00:45:15before he became the punching bag for
- 00:45:17spoiling the 2000 election was actually
- 00:45:20a very important leader for civil
- 00:45:22society because he was the face of the
- 00:45:25consumer protection movement that that
- 00:45:27was so important in introducing all
- 00:45:28these pharmaceutical transport and and
- 00:45:31other regulations so that sort of Civil
- 00:45:34Society movement is going to be very
- 00:45:35important and the issue is actually not
- 00:45:38new at all in the first few years where
- 00:45:42personal computers were coming on board
- 00:45:44there were people like this gentleman
- 00:45:45Ted Nelson who thought who outlin a very
- 00:45:49different vision of personal computers
- 00:45:52Ted Nelson and people like him thought
- 00:45:54that the problem with information was
- 00:45:57that companies like IBM were controlling
- 00:45:59it and the personal computers were the
- 00:46:01tool that would enable the
- 00:46:02democratization of information both in
- 00:46:04people's private lives and in the
- 00:46:06production process but of course who
- 00:46:09controlled the personal computers it was
- 00:46:11you know Microsoft and other companies
- 00:46:14that created large large companies that
- 00:46:16created tools for other large companies
- 00:46:18so the Ted Nelson hope never
- 00:46:22materialized but I think perhaps in the
- 00:46:23age of AI we can do
- 00:46:25better and and the last thing I'm going
- 00:46:27to show you
- 00:46:30is that this is
- 00:46:32not completely unprecedented either so
- 00:46:36it's not a complete pipe dream to say
- 00:46:39Civil
- 00:46:40Society perhaps government
- 00:46:42regulation Democratic input could then
- 00:46:46shape the direction of
- 00:46:48technology in a more beneficial
- 00:46:50direction again we could go back to the
- 00:46:52British industrial re second phase of
- 00:46:53the British Industrial Revolution after
- 00:46:541840s 1850s but here's a more recent
- 00:46:57example
- 00:46:58energy the energy
- 00:47:01sector you know okay fine we're not
- 00:47:04doing all that much in climate change
- 00:47:05about climate change we're probably
- 00:47:07going to exceed two and a half% two and
- 00:47:092.5 degrees Centigrade over
- 00:47:10preindustrial times but today actually
- 00:47:13the world is much better than it was 20
- 00:47:15years ago because 20 years ago none of
- 00:47:17the renewable Technologies were even
- 00:47:19close to being cost competitive in
- 00:47:22electricity production they were about
- 00:47:2310 times as expensive and starting in
- 00:47:27the late
- 00:47:292000s there's a complete Decline and now
- 00:47:32for electricity
- 00:47:35production onshore wind offshore wind
- 00:47:37and different types of Solar
- 00:47:38Technologies are all cost competitive
- 00:47:40with fossil fuels how did that happen it
- 00:47:42happened first of all because of
- 00:47:44innovation there was a huge number of
- 00:47:46patents and great exploitation of
- 00:47:49economies of scale in solar panel
- 00:47:51production and some other tool
- 00:47:52production why did it happen well
- 00:47:55because there was pressure from Civil
- 00:47:57Society a few countries introduced uh
- 00:48:00carbon taxes some countries including
- 00:48:03starting in California introduced
- 00:48:04regulations that made fossil fuels less
- 00:48:06profitable and many countries including
- 00:48:09the US started providing subsidies to
- 00:48:11the more socially beneficial direction
- 00:48:13of research which means Renewables
- 00:48:15rather than fossil fuels so what
- 00:48:18happened in energy I think can happen in
- 00:48:20how we use Ai and I I think the past
- 00:48:24shows that automatic hope that new
- 00:48:27technologies will naturally benefit
- 00:48:29democracy will naturally benefit workers
- 00:48:31will naturally benefit all of us is a
- 00:48:33little bit too much wishful thinking but
- 00:48:35the potential is there and if we create
- 00:48:38the right Power Balance and the right
- 00:48:40direction for technology we have more
- 00:48:42hope thank
- 00:48:53you thank you Darren so now we're moving
- 00:48:55into the the Q and
- 00:48:57portion and we've got about 20 25
- 00:48:59minutes um to uh to hear from all of you
- 00:49:04and uh and and hear more from from
- 00:49:06Darren so uh as I mentioned at the start
- 00:49:09if you want to post some questions uh uh
- 00:49:13essentially anonymously you're you're
- 00:49:14very welcome to do so uh using that QR
- 00:49:18code from uh from your phone um but
- 00:49:21we've also got microphones around the
- 00:49:22audience so so maybe um we start with an
- 00:49:25audience question if I can can see any
- 00:49:27any
- 00:49:28hands okay so I don't know where the
- 00:49:30microphones are though so oh here we go
- 00:49:33so let's let's start over here because
- 00:49:34it's probably easiest for you to get
- 00:49:36to and uh and we'll go from there uh hi
- 00:49:40thanks very much that was really
- 00:49:41insightful uh I was wondering about
- 00:49:43power not only in an ideological sense
- 00:49:46but also in the sense of Market power
- 00:49:48and whether the emergence of new
- 00:49:50technologies and the r in productivity
- 00:49:52may be associated with the emergence of
- 00:49:55Superstar firms and consequently
- 00:49:57concentrated labor markets may actually
- 00:49:59lead to a depression in wages I wonder
- 00:50:01if you think that's a realistic concern
- 00:50:03and Absolut should be consider
- 00:50:04absolutely uh thank you very much for
- 00:50:05that question and in fact there is a
- 00:50:08third reason why technological change
- 00:50:10may not
- 00:50:13generate prosperity for regular people
- 00:50:16or workers and it is related to Market
- 00:50:18power so if you have a new techn so you
- 00:50:20know when we're talking of prosperity
- 00:50:22here we're talking of real wages wages
- 00:50:24divided by some sort of priceing de so
- 00:50:27imagine we have a new technology but at
- 00:50:29the same time it increases the market
- 00:50:30power of a few companies so they start
- 00:50:32charging higher markups higher prices so
- 00:50:34then the price declines that would have
- 00:50:36led to the real wage increase would not
- 00:50:39happen and I think that is a concern and
- 00:50:41indeed as you said there are many
- 00:50:43sectors where we are seeing greater
- 00:50:46concentration the reason why I have not
- 00:50:47emphasized it is because emphasizing two
- 00:50:50rather than three is a little simpler
- 00:50:51but also because many of the tech
- 00:50:54sectors are not monetizing
- 00:50:57their products via higher prices and
- 00:51:00that's part of the Paradox of antitrust
- 00:51:02today you know uh there are people like
- 00:51:05Lina Khan and Jonathan canther in the
- 00:51:07United States who are really worrying
- 00:51:09about these things but there's a big
- 00:51:10barrier against them which is the usual
- 00:51:13antitrust argument would be well you
- 00:51:15know you've come to Corner the market
- 00:51:17you're charging higher prices well
- 00:51:18Facebook doesn't change any prices it
- 00:51:20just takes your data and monetizes it
- 00:51:22Google doesn't charge any prices Amazon
- 00:51:24actually well you know it's charges
- 00:51:27prices but often times it cuts prices
- 00:51:29relative to competitors because they're
- 00:51:30monetizing it differently they're also
- 00:51:33the Venture Capital based model of grow
- 00:51:36very fast is creating a long range of
- 00:51:39long time period in which these
- 00:51:40companies are not really charging big
- 00:51:42markups so I think it's important but
- 00:51:44that's why I haven't emphasized that as
- 00:51:46much but on the other hand I would say
- 00:51:49that business model of monetizing data
- 00:51:51actually really is pricious because it
- 00:51:54closes the door or it makes it hard
- 00:51:56harder for the types of tools that I was
- 00:51:59talking about that would make workers
- 00:52:01more productive why because if the main
- 00:52:03way of making money is collect data from
- 00:52:05people and monetize that through digital
- 00:52:07ads then you don't put your energy into
- 00:52:09finding better ways of making workers
- 00:52:11more
- 00:52:12productive so let me let me come to a
- 00:52:14question that's related that I've got
- 00:52:17here and um although just to to tea up
- 00:52:19the next question hands up okay maybe
- 00:52:22we'll take some maybe down here next but
- 00:52:25but this is I think somewhat related um
- 00:52:28someone on on on here is asking or
- 00:52:30suggesting that maybe um that early on
- 00:52:34benefits of Technology are more
- 00:52:36concentrated um but maybe later on and
- 00:52:39they're suggesting maybe it's a few
- 00:52:41decades but I don't know um you know
- 00:52:44there's a delay and then that spreads
- 00:52:46what what what do you think well that's
- 00:52:47a great question it's a very very
- 00:52:48important point so the question is
- 00:52:51obviously the British Industrial
- 00:52:53Revolution narrative that I provided
- 00:52:54suggests that you know early on things
- 00:52:56didn't work out and later they worked
- 00:52:58out but I think the question is was that
- 00:52:59automatic yeah or was that the result of
- 00:53:02some difficult choices difficult
- 00:53:04institutional adjustments so we can't
- 00:53:06know that we don't have the
- 00:53:07counterfactual history but
- 00:53:09my guess is that a if we did not create
- 00:53:13the trade unions if the trade unions
- 00:53:15they were persecuted as heavily in the
- 00:53:17beginning of the 20th century in the UK
- 00:53:19as they were in the beginning of the
- 00:53:2019th century if UK did not democratize
- 00:53:24and if the emphasis remained on
- 00:53:26automation it would have happened even
- 00:53:28more slowly and you know 90 years is
- 00:53:31already pretty slow that's three
- 00:53:32generations so my argument is that we
- 00:53:35can do it much faster with the right
- 00:53:37technological choices and right
- 00:53:38institutional choices so I wouldn't say
- 00:53:40it's automatic but sure there are many
- 00:53:43adjustments that do take place Norms
- 00:53:44change institutions change uh technology
- 00:53:46changes and there are Market processes
- 00:53:48that work more slowly absolutely I think
- 00:53:50I'm not ruling those out but I think the
- 00:53:52automatic is not the just the the only
- 00:53:55order of the day here here okay question
- 00:53:58over here
- 00:53:59please sorry thank you uh uh Professor
- 00:54:04uhu um it's it's more not question but
- 00:54:07more like I wonder how uh this kind of
- 00:54:11uh British revolution uh Industrial
- 00:54:13Revolution situation is leading us to
- 00:54:15like create a fear mongering of somebody
- 00:54:18or something stealing away our job like
- 00:54:21in the past people were so afraid that
- 00:54:23their job is going to be taken away by
- 00:54:24machine and now it's people are afraid
- 00:54:27again that it's going to be taken away
- 00:54:28by the AI my my comment is uh Professor
- 00:54:32is that uh in the past people have to
- 00:54:36use uh their muscle to work for the
- 00:54:38betterment of themselves and then it it
- 00:54:41evolves into the usage of brain and
- 00:54:43maybe in the future it's going to be
- 00:54:45involved into something else that is
- 00:54:46like perhaps we no longer need to be
- 00:54:49busy about doing something clal for
- 00:54:50example that is do uh is going to be
- 00:54:52taken away by the AI but then human
- 00:54:55Humanity in the future could be more
- 00:54:57focused on something that is more useful
- 00:54:59or more um using the heart instead of
- 00:55:02the brain if if if you get my uh Point
- 00:55:05uh my sense yeah I I get your point and
- 00:55:08and but but but I partially disagree uh
- 00:55:11first of all I mean I think you know
- 00:55:14there's a thin line separating concerns
- 00:55:17and fear mongering so I would say
- 00:55:19concerns are different than fear
- 00:55:21mongering if we are also seeing a path
- 00:55:26that would avoid those concerns so there
- 00:55:29have been indeed you are 100% right
- 00:55:31periodic concerns about jobs
- 00:55:34disappearing but they haven't all been
- 00:55:36wrong some of them have been right some
- 00:55:37of them have been wrong so the evidence
- 00:55:40that I did not show but it's my uh my
- 00:55:42work with Pascal Restrepo for
- 00:55:45example uh when we looked at the
- 00:55:48introduction of robots which was the
- 00:55:50previous big technology that uh that uh
- 00:55:54people were very afraid of well they
- 00:55:56were right to be afraid of in the United
- 00:55:58States we find that for every one robot
- 00:56:00there were six jobs that were destroyed
- 00:56:02and uh and and net and uh and in local
- 00:56:06labor markets where robots were
- 00:56:08introduced there were significantly
- 00:56:09lower wages especially for workers who
- 00:56:10were engaged in manual
- 00:56:12tasks and uh but it's not in inevitable
- 00:56:16so in Germany when they introduced
- 00:56:18robots what they did at the same time
- 00:56:20especially in work in companies which
- 00:56:22had unions and work councils is that
- 00:56:25they also at the same time upgraded the
- 00:56:27work jobs so the workers who used to do
- 00:56:29say painting or welding now became
- 00:56:31technical workers and as a result the
- 00:56:33effects were not as negative so they
- 00:56:35were right workers were right to be
- 00:56:37concerned but there was another option
- 00:56:39now I think in my mind the fear
- 00:56:41mongering is when we're talking of
- 00:56:43artificial general intelligence and
- 00:56:44Killer Robots that's fear mongering
- 00:56:46because it doesn't give us much choice
- 00:56:47and it's a very by you know uh uh uh d
- 00:56:52false dichotomy you know either robots
- 00:56:55are either super super artificial
- 00:56:58intelligence is going to be great for us
- 00:56:59or it's going to kill us all so I think
- 00:57:01the issue is in the middle it's going to
- 00:57:03do nether but it's how we actually deal
- 00:57:06with it and you are absolutely right
- 00:57:08there is there was a hope early on that
- 00:57:12machines would take manual Works manual
- 00:57:15jobs heavy jobs and they will leave us
- 00:57:17with
- 00:57:18more uh enjoyable more satisfying and
- 00:57:22less dangerous jobs and some of that has
- 00:57:24happened in workplaces where robots have
- 00:57:27been introduced in the US and in Europe
- 00:57:29yes in the US I told you jobs have
- 00:57:31disappeared but remaining jobs became
- 00:57:33safer and we definitely don't want to go
- 00:57:35to the time where people carry things by
- 00:57:37hands when there are cranes and uh uh
- 00:57:39and and and and and automated machinery
- 00:57:42for carrying but the problem with AI is
- 00:57:44actually that it's not taking the the
- 00:57:47jobs that are more mundane you know if
- 00:57:49you look at what are the jobs that are
- 00:57:51affected by AI you know the safest jobs
- 00:57:53are custodial Protective Services
- 00:57:55constru ruction workers it's the office
- 00:57:58jobs it's some jobs that are semi
- 00:58:00creative that could be taken away by AI
- 00:58:02if we just go down the automation path
- 00:58:05so so it's actually and then yeah of
- 00:58:07course perhaps in 100 years time we can
- 00:58:10have other adjustments but but again
- 00:58:12it's the automatic versus non automatic
- 00:58:14issue so I come to another poster
- 00:58:17question um really again I think an
- 00:58:19inequality around access to AI um and so
- 00:58:23you know we might all be sitting here
- 00:58:25and it's great we can access CH GPT or
- 00:58:28co-pilot or Claude or gemini or whatever
- 00:58:30it is for whatever purpose great but
- 00:58:33that's not true everywhere around the
- 00:58:35world so the question is uh how will the
- 00:58:37introduction of AI influence the
- 00:58:38economic landscape of regions with
- 00:58:40limited technological access
- 00:58:42particularly in contrast to the
- 00:58:43advancements seen in more affluent
- 00:58:45Nations absolutely that's very important
- 00:58:46and that's what that's the thing I very
- 00:58:48briefly hinted at which is what about
- 00:58:50the developing World developing world
- 00:58:52doesn't have access even if they have
- 00:58:53access the global division of labor is
- 00:58:55going to change in massive way you know
- 00:58:57what are the you know look at the
- 00:59:00globalization experience of the uh 18th
- 00:59:04century India became hugely
- 00:59:07de-industrialized that was like the uh
- 00:59:10place that was most advanced in textiles
- 00:59:13globalization became uh very damaging to
- 00:59:16India well look at the globalization in
- 00:59:19the 1960s 1970s well South Korea Taiwan
- 00:59:23later China did very well under
- 00:59:25globalization why globalization created
- 00:59:28opportunities that they didn't have in
- 00:59:30particular it enabled them to specialize
- 00:59:32in things like textiles or uh or
- 00:59:34furniture or toys that were low skill
- 00:59:38low technology sectors that could be an
- 00:59:41engine of growth well the question is is
- 00:59:43AI going to create that sort of thing
- 00:59:45and it doesn't seem like that quite the
- 00:59:47opposite some of the services that could
- 00:59:49be done in the in the developing world
- 00:59:51would actually be completely reshored
- 00:59:53with AI so I think the complic
- 00:59:55implications are are going to be uh more
- 00:59:58complex but even having access to
- 01:00:01Ai and knowing how to use it may not be
- 01:00:04enough so imagine that we have you know
- 01:00:07uh uh let me give you an example from
- 01:00:12journalism so we have 100,000
- 01:00:14journalists say in the United States
- 01:00:16that's not not quite that many but okay
- 01:00:19but we we teach all of them to be very
- 01:00:21good at prompt
- 01:00:22engineering but if ai go in the
- 01:00:26automation Direction and all that it can
- 01:00:28do for journalists is that more of their
- 01:00:30tasks can be completed doesn't matter
- 01:00:32whether you're very good at prompt
- 01:00:34engineering what's going to happen is
- 01:00:35that we're not going to need 100,000 but
- 01:00:36we're going to need not only 20,000
- 01:00:38journalists so journalists are not going
- 01:00:40to fare very well what would it take for
- 01:00:43them to Fair better well if instead of
- 01:00:44just automating if we create AI if we
- 01:00:47develop AI such that we can actually
- 01:00:49provide better information to journalist
- 01:00:50so that they can do more sophisticated
- 01:00:52research so that's really the nature of
- 01:00:55Technology whether you get you know how
- 01:00:56to use the technology or not is
- 01:00:58important but it's not just the only
- 01:01:00determining Factor okay another question
- 01:01:02from the room how about we go um I
- 01:01:05already have a microphone okay whoever's
- 01:01:07got a microphone I can't see where you
- 01:01:08are hi oh there you are fantastic um hi
- 01:01:11my name is Julia I'm an undergraduate
- 01:01:12here studying economics um my question
- 01:01:15relates to a different type of
- 01:01:16inequality you show a graph with the
- 01:01:19Divergence in Ro wages like related to
- 01:01:22education um and noticeably there's less
- 01:01:25Divergence among women women linking
- 01:01:27that in with what you say about task
- 01:01:29replacement um is the implication that
- 01:01:32tasks typically perform by women are
- 01:01:34being replaced to a lesser extent or how
- 01:01:36would you explain that di absolutely so
- 01:01:38essentially 100% you're very very
- 01:01:40perceptive that if you look at the
- 01:01:43period that I focused on which is where
- 01:01:45some ofice work was being automated with
- 01:01:47Software System but a lot of automation
- 01:01:49was taking place for blue color jobs
- 01:01:52that was on more the onus was heavier on
- 01:01:55men and in fact you're 100% right pretty
- 01:01:59much every dimension of
- 01:02:01inequality increase except by gender the
- 01:02:04gender
- 01:02:04inequality narrowed for other reasons
- 01:02:07but the automation of male jobs
- 01:02:09especially male jobs that paid High
- 01:02:11wages help but here is the bad news I
- 01:02:14have recently done the study of AI and
- 01:02:17who might be you know it's very you know
- 01:02:21speculative because we don't it hasn't
- 01:02:22happened yet it's the future but if you
- 01:02:24look at the types of jobs that generate
- 01:02:26AI can perform they're more female jobs
- 01:02:28so it looks like now generative AI might
- 01:02:32act towards increasing the gender gap
- 01:02:34rather than the earlier automation that
- 01:02:36might have a little bit closed
- 01:02:38it okay I got another one from the room
- 01:02:42I know maybe go to
- 01:02:46you thank you and please excuse my voice
- 01:02:49it's about to be gone but um I wanted to
- 01:02:52ask you about that future where you have
- 01:02:54that one person who's who watching the
- 01:02:56dog you know and you don't know you
- 01:02:58don't have all these other workers how
- 01:03:00does the the notion of a universal basic
- 01:03:02wage fit into that and what would be
- 01:03:04differences between countries yeah so
- 01:03:07excellent question and thank you for
- 01:03:08bringing it up you know because there's
- 01:03:12like a big gaping hole in what I talked
- 01:03:14about which is okay fine why do we need
- 01:03:17wages to create shared Prosperity let
- 01:03:20Elon Musk earn everything okay some
- 01:03:22outman earns some too and then we
- 01:03:24redistribute all of it
- 01:03:26and then you know you can choose your
- 01:03:28favorite method of redistribution could
- 01:03:29be Ubi it could be something else but we
- 01:03:32have a huge level of
- 01:03:34inequality and then we just use the
- 01:03:36fiscal system to redistribute it what's
- 01:03:39wrong with that well I would say there
- 01:03:41are three things wrong with that first
- 01:03:43of all I'm suggesting an alternative
- 01:03:45which is redirecting technology so that
- 01:03:47we generate more equal distribution of
- 01:03:48income so saying that we're going to go
- 01:03:51to Ubi is defe in my opinion because it
- 01:03:54says we can't do anything else there is
- 01:03:56nothing we can do to change technology
- 01:03:58or technology is
- 01:04:00unchangeable the future is one of only
- 01:04:02Elon Musk some Alman and Mark Zuckerberg
- 01:04:05making money were going to be all
- 01:04:07dispensable well that's a sad type of
- 01:04:10defeatism second I don't think po
- 01:04:13political economy of it doesn't work I
- 01:04:15haven't seen many billionaires who are
- 01:04:17willingly giving their money today I
- 01:04:18mean the only thing that they are
- 01:04:19willing to do is Charity which often is
- 01:04:24a way of f further boosting their
- 01:04:27Prestige and
- 01:04:28power so what makes you think that
- 01:04:33tomorrow suddenly they'll say we earn
- 01:04:34everything but we're going to distribute
- 01:04:36the rest of it well perhaps if they're
- 01:04:37really worried about a revolution
- 01:04:39perhaps but it's not going to be that
- 01:04:41easy but the third one is the one that
- 01:04:43would really worry me and that again
- 01:04:45takes us from the realm of pure
- 01:04:48economics to more broader social
- 01:04:50considerations if we create a world or
- 01:04:52if we are in the world in which say 5%
- 01:04:57of the population does the earnings the
- 01:05:00remaining 95% is completely useless they
- 01:05:02don't have any jobs they can't do
- 01:05:05anything they just live off the crumbs
- 01:05:07that's going to be a hugely hierarchical
- 01:05:09hugely dystopian world with status gaps
- 01:05:13that are just massive even relative to
- 01:05:15what we are seeing today so I don't
- 01:05:16think that's going to be a happy world
- 01:05:18so yes if indeed either because I'm
- 01:05:21wrong and it's inevitable that we go to
- 01:05:23massive inequality or it's not
- 01:05:25inevitable but we don't make the right
- 01:05:27policy choices yes of course we have to
- 01:05:29create some sort of way of
- 01:05:30redistributing them but hopefully it
- 01:05:32doesn't come to that thank you for
- 01:05:33raising that let's do another one from
- 01:05:36the room okay can we go down the front
- 01:05:39here
- 01:05:44please I think both of you had your hand
- 01:05:46up so you can find it out it's up to you
- 01:05:49thank you um thank you for your talk um
- 01:05:53I want to ask that um as we have Cod in
- 01:05:55your talk Milton Freedman um clearly
- 01:05:59thinks that market mechanism is a better
- 01:06:01arbitrator of um Power Balance and stuff
- 01:06:05than political channels but also I feel
- 01:06:08that there are also problems with um
- 01:06:11politics because uh sorry market
- 01:06:13mechanism because uh clearly economic
- 01:06:16power is very approximate to political
- 01:06:18power and um as we see um more and more
- 01:06:22technology techn attack Giants Rising
- 01:06:25maybe the legislations that um are
- 01:06:28supposed to curb their power are like
- 01:06:30received the most influence from them so
- 01:06:33um who do you think is the better
- 01:06:36arbitrator of this power B well I don't
- 01:06:38think there is a perfect arbitrator you
- 01:06:40you know absolutely I I think there's
- 01:06:42sometimes in discussions and sometimes
- 01:06:45in economics debates there is this sort
- 01:06:49of rhetoric that somehow politics is
- 01:06:53something that comes in addition
- 01:06:56like on top like things are working well
- 01:06:59if only politics wasn't part of it and I
- 01:07:01think that's not the right way to think
- 01:07:03about it politics is with us all the
- 01:07:04time politics power relations are always
- 01:07:07there in workplaces in markets in social
- 01:07:10situations and it is a balance between
- 01:07:12the market and politics and exactly like
- 01:07:14you said if somebody has a lot of
- 01:07:15economic power they will tend to have a
- 01:07:17lot of political power so saying oh the
- 01:07:20Market's going to sort out things if we
- 01:07:22just leave politics out is is a fantasy
- 01:07:26but the political process is very
- 01:07:27unequal sometimes much more unequal than
- 01:07:29the economic process so that's why we
- 01:07:31have to try to build the right sort of
- 01:07:33Institutions and Norms to balance them
- 01:07:35out but it's not going to be perfect and
- 01:07:37when unions were
- 01:07:39strong they
- 01:07:41helped curb some abuses of workers but
- 01:07:44then the workers
- 01:07:46sometimes had too much power in some
- 01:07:48workplaces and they use that power
- 01:07:51incorrectly or badly like resisting
- 01:07:53technological change in the British
- 01:07:55printing in the British printing
- 01:07:57industry uh unions resisted printing
- 01:08:00machines making the industry fall behind
- 01:08:0320 years so any group having too much
- 01:08:06power is going to be uh a recipe for
- 01:08:10certain types
- 01:08:12of problematic outcomes so some sort of
- 01:08:15balance if we can achieve it would be
- 01:08:17the thing to strive for okay so let me
- 01:08:22let me go with another question that
- 01:08:23somebody's put uh online here um this is
- 01:08:27this getting to what you're talking
- 01:08:28about before around essentially you know
- 01:08:30one one view of AI is substituting
- 01:08:32humans versus the other one
- 01:08:34complimenting or or augmenting um and
- 01:08:36and the question is really just about
- 01:08:38well what what skills do you think will
- 01:08:39see we will see complimented versus
- 01:08:41substituted by Ai and I guess really
- 01:08:44they're getting out is how do we then
- 01:08:46think about that in the context of
- 01:08:47directing or redirecting technological
- 01:08:50development into sort of the the right
- 01:08:53way shall we say as opposed to the other
- 01:08:54perfect there a great great question but
- 01:08:56my answer is it will depend because it
- 01:08:58really depends on how we develop AI so
- 01:09:00here is one skill that I would love to
- 01:09:03see as complement to AI
- 01:09:07electricians
- 01:09:09so in the United Kingdom in the US we
- 01:09:12have a shortage of electricians that's
- 01:09:13going to get only worse with the
- 01:09:14electrification of the
- 01:09:16grid what is the problem the problem is
- 01:09:18we're not training enough electricians
- 01:09:20and the electricians are required to
- 01:09:22deal with more and more complex problems
- 01:09:27so is that a substitute or compliments
- 01:09:29to AI well you can say well we can try
- 01:09:31to automate these jobs in which case
- 01:09:32they would be substitutes but actually
- 01:09:34the better way would be use AI to
- 01:09:36provide better information and better
- 01:09:38training opportunities so that they can
- 01:09:39upgrade their skills they can recognize
- 01:09:41problems and diagnose problems and
- 01:09:43perform more sophisticated tasks in
- 01:09:45which case AI would be a compliment to
- 01:09:47electricians same thing for plumbers
- 01:09:49same things to other manual workers so I
- 01:09:51think there's a lot of scope for
- 01:09:52creating those complementarities but it
- 01:09:54would require the technology to change
- 01:09:56in that
- 01:09:58direction okay I think we got time for
- 01:10:00maybe one more question we'll take that
- 01:10:02from the room um so let's go in the very
- 01:10:06back because I've had a front bias here
- 01:10:08so let's go up the back there um and
- 01:10:11apologies to the microphone people
- 01:10:13having to walk up the
- 01:10:17stairs very interesting thank you for
- 01:10:19your time I'm studying artificial
- 01:10:21intelligence for business here um the
- 01:10:24question is
- 01:10:26you're talking you're talking about the
- 01:10:29the the capabilities that are
- 01:10:32complimenting humans through Ai and the
- 01:10:36and the ones that replace uh and you
- 01:10:39talk about Powers corporate Powers
- 01:10:42industrial Powers I think there is a
- 01:10:44double narrative at the same time
- 01:10:47happening because when you see
- 01:10:50yesterday H the LinkedIn and Microsoft
- 01:10:53work report telling that 7 5% of white
- 01:10:56colors are using AI in in the United
- 01:11:00States but at the same time h two thirds
- 01:11:03of the companies in the United States
- 01:11:05are developing projects to replace some
- 01:11:10of their Workforce through
- 01:11:12AI I my question is in this double
- 01:11:18narrative ER from your point of view
- 01:11:20which one is more uh
- 01:11:23realistic the AI complementing
- 01:11:26capabilities of humans or the companies
- 01:11:29making Replacements well that's a great
- 01:11:31question and that's the key question and
- 01:11:32I don't think I can make a forecast
- 01:11:36because it's not like the weather it's
- 01:11:38something we're going to choose
- 01:11:40so if Microsoft and open Ai and Google
- 01:11:45decide it's substituting for workers
- 01:11:48then we're going to end up with
- 01:11:50substituting for workers unless some
- 01:11:52other companies come up so their choices
- 01:11:54are going to matter
- 01:11:55but also I think hype about AI is not
- 01:11:59helping and and the number that you just
- 01:12:02quoted is pure hype you know
- 01:12:06uh very few workers right now are really
- 01:12:09using AI in the United States so you can
- 01:12:11count many workers using AI because if
- 01:12:13you're using Microsoft Word Microsoft
- 01:12:14Word has a co-pilot the co-pilot has AI
- 01:12:17in it so then you can say you do but I
- 01:12:20uh arranged a big with my collaborators
- 01:12:24big survey for all of businesses in the
- 01:12:27United States through the Census Bureau
- 01:12:29and in 2020 1.5% of us businesses had
- 01:12:34any investment in AI okay perhaps it has
- 01:12:36increased a little bit in the last three
- 01:12:38years but it's still very very much
- 01:12:41embryonic so it's all to play for it's
- 01:12:43going to be the next decade or so where
- 01:12:45we're going to see how much AI what type
- 01:12:47of AI and who's going to benefit from it
- 01:12:49so I think being realistic about where
- 01:12:52we are and what we can do is a very
- 01:12:54important part of it thank you okay and
- 01:12:58that's a really good point to end on so
- 01:13:00I want to say a massive thank you to to
- 01:13:02Darren for joining us and and giving us
- 01:13:04your thoughts and thank you to all of
- 01:13:05you for coming along as well thank you
- tehnologie
- economia
- automatizare
- AI
- inegalitate
- putere
- progres economic
- revoluție industrială
- istorie
- inovație